Immune Complexes Induce TNF-Α and BAFF Production from U937 Cells by HMGB1 and RAGE Tory Mediators8,9

European Review for Medical and Pharmacological Sciences 2017; 21: 1810-1819

Immune complexes induce TNF-a and BAFF production from U937 cells by HMGB1 and RAGE

X.-J. GAO1,2, Y.-Y. QU2, X.-W. LIU2, M. ZHU2, C.-Y. MA2, Y.-L. JIAO2, B. CUI2, Z.-J. CHEN3, Y.-R. ZHAO2

1Center of Laboratory Medicine, The Affiliated Yantai Yuhuangding Hospital of Qingdao University, Yantai, China 2Central Laboratory, Provincial Hospital Affiliated to Shandong University, Jinan, China 3Research Center for Reproductive Medicine, Provincial Hospital Affiliated to Shandong University, Jinan, China

Abstract. – OBJECTIVE: This study investi- Key Words: gated the effects of immune complexes (ICs) on Immune complexes, U937 cells, High mobility group tumor necrosis factor α (TNF-α) and B cell-ac- box protein 1, Receptor for advanced glycation end tivating factor (BAFF) production from U937 products, Systemic lupus erythematosus. cells and further explored the mechanism. MATERIALS AND METHODS: U937 cells were incubated with necrosis supernatant or systemic lupus erythematosus (SLE) sera alone, or Introduction their combination. The expression of TNF-α and BAFF was determined by Real-time poly- Systemic lupus erythematosus (SLE) is a mul- merase chain reaction and enzyme-linked im- ti-systemic involvement and chronic progressive munosorbent assay. High mobility group box protein 1(HMGB1) A-box was produced by gene autoimmune disorder characterized by production recombination. HMGB1 A-box and anti-receptor of autoantibodies against nucleic components such for advanced glycation end products (RAGE) as DNA and nucleosomes and formation of immu- antibody were adopted in the blocking experi- ne complexes (ICs) due to polyclonal B cell acti- ments. The importance of DNA for cytokine in- vation1,2. Being inactive in its free form, however, duction was investigated by DNase treatment. mammalian DNA in the form of ICs has potent RESULTS: The combination of necrosis su- immunostimulatory properties3,4. The necessity for pernatant and SLE sera induced the expression of TNF-α and BAFF significantly increased complex formation suggests that endogenous DNA compared to necrosis supernatant or SLE se- must exist in association with another moiety or ra alone. Recombinant HMGB1 A-box protein structure to enable access to the corresponding re- was purified, and TNF-α and BAFF production, ceptor. High mobility group box protein 1 (HMGB1) which were induced by this combination, was is known as a non-histone nuclear protein that binds blocked via HMGB1 A-box and anti-RAGE anti- DNA in a non-sequence-specific manner, modi- body. Moreover, we found that DNA component fying chromosomal architecture and facilitating is important for the immunostimulatory activity 5 of this combination. gene transcription . HMGB1 has been highlighted CONCLUSIONS: ICs containing DNA can pro- as a pro-inflammatory mediator when it is released mote TNF-α and BAFF production in U937 cells, from cells. Binding of HMGB1 to the receptor for and this process can be mediated by HMGB1 advanced glycation end products (RAGE), Toll-like and RAGE. One possible mechanism of increas- receptor-2 (TLR-2) or TLR-4 can lead to the release ing BAFF production in SLE is proposed in this of pro-inflammatory cytokines6,7. However, highly study whereby B cell activation, antibody pro- purified HMGB1 has a weak or no pro-inflamma- duction and ICs stimulated monocytes may create a vicious cycle that leads to B cell hyper- tory activity; for full activity, it may need to form a activity, which can be of importance for SLE complex with another component to activate inflam- etiopathogenesis. mation by enhancing effects of other pro-inflamma-

1810 Corresponding Author: Yueran Zhao, MD; e-mail: [email protected] Immune complexes induce TNF-α and BAFF production from U937 cells by HMGB1 and RAGE tory mediators8,9. Of note, HMGB1 can bind avidly demonstrated that ICs containing DNA could pro- to immunostimulatory molecules such as lipopoly- mote the increase of TNF-α and BAFF production saccharide (LPS), DNA or interleukin (IL)-1β so as in U937 cells via HMGB1 and RAGE. Therefore, to promote their activity in a synergistic fashion9,10. the presented data may help to better understand the Given that increased DNA or nucleosomes in the cytokine disturbance and pathogenesis of SLE. blood of SLE has been attributed to secondary necrosis, it appears reasonable to expect that released nucleosomes or DNA may carry along the tightly Materials and Methods bound pro-inflammatory HMGB1. Urbonaviciute et al11 demonstrated that nucleosomes were complexes Culture of U937 Cells and Induction with HMGB1 in the supernatants of secondary ne- of Cell Necrosis crotic cells and HMGB1 was present in circulating The human monocyte line U937 (ATCC 1593) ICs from SLE patients. Monocyte is a key compo- was cultured in Modified Roswell Park Memorial nent of the innate immune system involved in the Institute (RPMI)-1640 medium (TransGen Biote- regulation of adaptive immune responses. Aberrant ch, Beijing, China) supplemented with 10% fetal activation of the adaptive immune system in SLE calf serum (FCS) (HyClone, Logan, UT, USA), is augmented through the dysregulated production L-glutamine (2 mM, MP Biomedicals, Shanghai, of cytokines1. Abnormally activated monocytes/ China), HEPES (20 mM, MP Biomedicals, Shan- macrophages are major sources of some cytokines ghai, China), penicillin (60 μg/ml, Solarbio, Bei- and are usually found in the sera and tissues from jing, China), and streptomycin (100 μg/ml, So- 12-14 patients with SLE , which may promote the acti- larbio, Beijing, China), at 37ºC in 5% CO2. The vation of auto-reactive T and B cells. Among the- induction of U937 cell necrosis was performed by se cytokines, tumor necrosis factor-alpha (TNF-α) repeat freeze thawing as previously described14. contributes to tissue inflammation and damage15, Cell debris was removed by centrifugation (400 and B cell can activate factor from the TNF family g for 10 min, GTR22-1, Beijing Era Beili Cen- (BAFF) that promotes the survival and proliferation trifuge Co., Ltd, Beijing, China) and necrotic cell of B lymphocytes7,16. BAFF, also known as B lym- supernatant was isolated and stored in -80°C for phocyte stimulator (BLyS), is mainly produced by subsequent preparation of ICs. monocytes. Circulating BAFF levels are elevated and correlate positively with anti-dsDNA antibody Serum Preparation titers and disease activity in SLE patients17,18. Huma- SLE sera were prepared from the blood samples nized anti-BAFF antibody and BAFF decoy recep- of SLE patients in our hospital who fulfilled the tor significantly reduced the level of autoantibodies American College of Rheumatology (ACR) 1997 and B-cell activation in SLE19. Thus, BAFF is likely revised criteria for the classification of SLE23. Ve- to play a critical role in the activation of antigen-dri- nous blood was collected from SLE patients during ven autoimmune B cells in human SLE. In SLE, one a flare or with active disease based on clinical clas- explanation for excessive cytokine production is the sification criteria into tubes without anticoagulants. presence of nucleic acid-containing ICs in the serum After isolated, sera were screened by anti-dsDNA of lupus patients. Recent reports13,20,21 have shown antibody ELISA test (Euroimmun, Lübeck, Ger- that nucleic acid-containing ICs are capable of in- many). Only sera with positive anti-dsDNA titer ducing TNF or IFN-α production from peripheral greater than 1:100, Shanghai KaLang Biologi- blood mononuclear cells (PBMCs), plasmacytoid cal Technology (Co., Ltd, Shanghai, China) were dendritic cells (PDCs) and B cells. Moreover, the in- selected and mixed. The mixed sera were passed teraction between HMGB1 and RAGE triggers the through a 0.45 μm filter and stored at -80°C. Also, activation of PDCs and B cells in response to lupus sera from healthy blood donors were prepared and ICs22. Given these findings, we suggest that ICs play used in some control experiments. All the samples their biologic roles primarily through stimulating were collected after informed written consent was monocyte activation. With many monocytic cha- obtained from all participants, and this study was racteristics, U937 cells may be chosen as a model to approved by Ethics Committee of Shandong Uni- investigate the effect of itself or environmental agen- versity. ts on monocytic function. The aim of this study was to investigate the effects of ICs on TNF-α and BAFF Preparation of Immune Cells (ICs) production from U937 cells and further explored Necrotic U937 cell supernatant and mixed SLE the mechanism for these effects. The results herein sera obtained above were used for preparation

1811 X.-J. Gao, Y.-Y. Qu, X.-W. Liu, M. Zhu, C.-Y. Ma, Y.-L. Jiao, B. Cui, Z.-J. Chen, Y.-R. Zhao of ICs. ICs were prepared according to previous TNF-α and BAFF mRNA Assays reports13,24. In this study, ICs were preformed by Total RNA was extracted using the TRIzol re- combining 0.5% (vol/vol) SLE sera and 12.5% agent (Invitrogen, Carlsbad, CA, USA). Then, 1 (vol/vol) necrosis supernatant in the complete μg of total RNA from each sample was used as a medium according to our preliminary experimen- template in the reverse transcription reaction to ts. The control groups were designed as follows: synthesize cDNA with random hexamer primers 12.5% necrosis supernatant or 0.5% SLE sera alo- according to Fermentas RT kit (Fermentas, Glen ne, the combination of 12.5% necrosis supernatant Burnie, MD, USA). All samples were reverse tran- and 0.5% normal sera, or 0.5% normal sera alone. scribed under the same condition (42ºC for 60 min) We found that the combination of 12.5% necrosis and from the same reverse transcription master mix supernatant and 0.5% normal sera did not modify to minimize differences in reverse transcription ef- the effect of necrosis supernatant alone, and 0.5% ficiency. The cDNA (2 µl) was amplified by Re- normal sera alone showed no stimulatory activity al-time PCR using a QuantiTect SYBR Green kit (data not shown). (TaKaRa, Otsu, Shiga, Japan) in ABI 7500 PRISM system (ABI, Foster City, CA, USA). The primers Expression and Purification of Human were synthesized and validated by Sangon Bio (Co. HMGB1 A-box Ltd., Shanghai, China) (Table I). The housekeeping Sequence of human HMGB1 A-box was op- gene, glyceraldehyde-3-phosphate- dehydrogenase timized with OptimumGeneTM for expression in (GAPDH) EarthOx (San Francisco, CA, USA), E. coli and synthesized (GenScript, Piscataway, was used as an internal control. The PCR was per- NJ, USA). Target sequence was ligated into a formed for 30 cycles of 95ºC for 15 s, 60ºC for 30 bacterial expression vector (pQE-T7-2; Qiagen, s and 72ºC for 40 s. The relative gene expression Hilden, Germany). After transformed by the iden- of TNF-α and BAFF was analyzed by the 2-∆∆Ct tified recombinant vector pQE-A-box, E. coli method15 and was depicted as the extent of change BL21 was induced by isopropyl β-D-1-thiogal- with respect to glyceraldheyde 3-phosphate dehy- actopyranoside (IPTG) (Boston Biomedical, Inc., drogenase (GAPDH) values. Boston, MA, USA) (1 mmol/L) for 4 h at 37ºC. The recombinant protein in bacterium lysates was TNF-α and BAFF Protein Assay analyzed by SDS-PAGE and identified by western Supernatants from U937 cultures were harve- blot using rabbit anti-human HMGB1 polyclonal sted after 24 h stimulation, and TNF-α and BAFF antibodies (Abcam, Cambridge, MA, USA). The protein levels were measured by sandwich ELI- recombinant protein was purified by Ni2+-NTA SA kits (R&D Systems, Minneapolis, MN, USA) (Sigma-Aldrich, St. Louis, MO, USA) affinity according to the manufacturer’s instructions. The chromatography and dialysis. Contaminating en- sensitivity for TNF-α and BAFF determination is dotoxin was removed by phase separation using 1.6 and 2.43 pg/ml, respectively. Triton X-114 as previously described25. Blocking test Cytokine Induction HMGB1 A-box was generated by gene recom- U937 cells were incubated at a density of 1×106 bination as described herein. Rabbit anti-human cells/ml in complete medium containing prefor- RAGE antibody was purchased from Abcam med ICs, necrosis supernatant or SLE sera alone (Cambridge, MA, USA). HMGB1 A-box and in 24-well flat-bottomed plates (Corning, New anti-RAGE antibody at different doses were re-

York, NY, USA) at 37°C and 5% CO2. The cel- spectively added to the U937 cultures before the ls were collected by centrifuging at 200 g for 5 addition of preformed ICs. min at 4°C into 1.5 ml Eppendorf centrifuge tubes (Beijing Biocoen Biotechnology Co., Ltd, Bei- Statistical Analysis jing, China). Also, DNase or RNase-treated ICs All analyses were based on three or more se- were used in some experiments. DNase I Roche parate experiments. Data were presented as mean Diagnostics product, Co., Ltd. (Shanghai, Chi- ± standard deviation. One-way ANOVA followed na) or RNase A (Sigma-Aldrich, St. Louis, MO, by the Bonferroni test for multiple groups or Stu- USA) of different concentrations was added to the dent’s t-test for two groups was used (p<0.05 was medium containing preformed ICs. After 1 h incu- considered to be significant). Statistical analyses bation at 37°C, the enzyme-treated ICs were used were performed by the SPSS16.0 (SPSS Inc., Chi- as described above. cago, IL, USA) statistical package.

1812 Immune complexes induce TNF-α and BAFF production from U937 cells by HMGB1 and RAGE

Figure 1. Necrosis supernatant combined with SLE sera activates TNF-α and BAFF production in U937 cells. (a) U937 cells were stimulated with 12.5% necrosis supernatant, 0.5% SLE sera or their combination. Cells were harvested at the indicated time points for RNA. The relative expression of TNF-α and BAFF was determined by Real-time PCR and was depicted as the extent of change on GAPDH values. (b) Expression of TNF-α and BAFF mRNA at the 8 h time point after stimulation. (c) ELISA performed on the supernatants of the stimulated cells to detect TNF-α and BAFF production at the 24 h time point.

Results activity of either necrosis supernatant or SLE sera alone. To determine the kinetics of cytokine ICs augment TNF-α and BAFF Expression expression, U937 cells were stimulated for 1, 4, in U937 Cells 8, 16, or 24 h with necrosis supernatant combi- Necrotic cells release significant quantities of ned with SLE sera, necrosis supernatant or SLE nuclear antigens26, which bind specific autoanti- sera alone. We detected maximal TNF-α mRNA bodies in SLE sera to form ICs. We investigated at about 4 h of stimulation, while BAFF mRNA whether the combination of necrosis supernatant expression peaked around 8 h (Figure 1a). The re- and SLE sera modified the immunostimulatory maining experiments in this study were based on

1813 X.-J. Gao, Y.-Y. Qu, X.-W. Liu, M. Zhu, C.-Y. Ma, Y.-L. Jiao, B. Cui, Z.-J. Chen, Y.-R. Zhao

Figure 2. The expression, identification and purification of recombinant human HMGB1 A-box. (a) SDS-PAGE of HMGB1 A-box. 1. pQE-A-box bacterium lysates before induction; 2. pQE-T7-2 bacterium lysates after inducing by IPTG for 4 h; 3-4: pQE-A-box bacterium lysates after inducing by IPTG for 4 h. (b) Western blot of HMGB1 A-box. 1, pQE-A-box bacterium lysates before induction; 2-3, pQE-A-box bacterium lysates after IPTG induction for 4 h. (c) SDS-PAGE of A-box purified by affinity chromatography on Ni2+-NTA column and dialysis. 1-2, purified recombinant A-box protein. Mr, relative molecular mass.

this data. The mRNA expression of both TNF-α te-polyacrylamide gel electrophoresis (SDS-PA- and BAFF in the group of necrosis supernatant GE) analysis showed that the relative molecular combined with SLE sera was significantly hi- mass (Mr) of the recombinant protein was about gher than that in untreated, necrosis supernatant 14 kD, which was consistent with the expected or SLE sera alone group (p<0.05 for TNF-α and (Figure 2a). We identified the recombinant protein p<0.01 for BAFF) (Figure 1b). Higher TNF-α as HMGB1 A-box by Western blot (Figure 2b). mRNA expression was observed in necrosis su- Then, by Ni2+-NTA affinity chromatography and pernatant group compared to the untreated group, dialysis, we obtained purified recombinant human but no statistical difference was reached. Also, no HMGB1 A-box protein (Figure 2c). significant difference in BAFF mRNA expression was found between necrosis supernatant and Antagonizing HMGB1 or Blocking RAGE untreated groups. SLE sera alone showed no si- Inhibits the Stimulatory Activity of ICs gnificantly stimulating effect in both TNF-α and To further investigate whether the stimulatory BAFF mRNA expression. To demonstrate that activity of necrosis supernatant combined with the increases in TNF-α and BAFF mRNA levels SLE sera on monocytic U937 cells was initiated were accompanied by protein production, ELISA by HMGB1 and RAGE, we adopted the blocking was performed on the supernatants of U937 cel- reagent, anti-RAGE antibody and recombinant ls stimulated with necrosis supernatant combined human HMGB1 A-box, which has been shown with SLE sera, necrosis supernatant or SLE sera to function as HMGB1 antagonist14. U937 cells alone for 24 h. The results were consistent with were treated with the combination (necrosis su- the mRNA data (Figure 1c). These data suggested pernatant and SLE sera) alone or in the presen- that the combination of necrosis supernatant and ce of anti-RAGE or A-box. We found that both SLE sera promoted a robust TNF-α or BAFF pro- HMGB1 A-box (20 μg/ml) and anti-RAGE (0.5 duction from U937 cells. μg/ml) antibody significantly reduced TNF-α mRNA expression induced by necrosis super- Identification and Purification of natant combined with SLE sera in U937 cells Recombinant Human HMGB1 A-box Protein (p<0.05) (Figure 3a). The TNF-α mRNA level We established the high-efficiency expression in the blocked ICs group was higher than that in system E. coli BL21-pQE-A-box by gene recom- the untreated group, but no statistical significance bination technique and induced the expression of was achieved. Moreover, we observed that either target protein. sodium Sodium dodecyl sulpha- anti-RAGE (0.5 μg/ml) or HMGB1 A-box (20 μg/

1814 Immune complexes induce TNF-α and BAFF production from U937 cells by HMGB1 and RAGE

BAFF-inducing activity (p<0.05) (Figure 4 a and b, respectively). However, RNase treatment was unable to lessen ICs-induced TNF-α and BAFF production in U937 cells (data not shown). Thus, these results demonstrated that the DNA component of ICs was important for inducing cellular activity.

Discussion

ICs can activate the immune cells to produce cytokines13,22,26. In this study we have investigated the possibility that ICs may stimulate the production of TNF-α and BAFF in SLE. We have demonstrated that ICs can significantly promote the expression of TNF-α and BAFF from human monocytic U937 cells. Furthermore, our data have provided evidence that HMGB1 can also induce ICs to mediate U937 activation through a mechanism involving RAGE, and the DNA component of ICs was necessary for its immunostimulatory activity. Recent reports26,27 have demonstrated that lupus ICs induce IFN-α and TNF production by PBMCs or PDCs. Here we observed that material from necrotic U937 cells, when Figure 3. HMGB1 A-box or anti-RAGE antibody inhib- combined with SLE sera, augmented TNF-α and its TNF-α and BAFF production from U937 cells induced BAFF production from U937 cells. Although ne- by preformed immune complexes (ICs). HMGB1 A-box or crotic cells were considered to be pro-inflamma- anti-RAGE antibody at indicated concentrations was added tory13, however, necrotic supernatant alone didn’t to the U937 cultures prior to the addition of preformed ICs. Cells were harvested at the 8 h time point for RNA. The rela- show significantly stimulatory activity in this tive expression depicted as the extent of change with respect study, which was consistent with a previous fin- to GAPDH values: (a) TNF-α and (b) BAFF. ding that necrotic material alone failed to induce IFN-α from PBMCs12. It wasn’t surprising that SLE sera alone were inactive, because the con- ml) significantly inhibited BAFF mRNA expres- centration of circulating ICs in serum was low sion triggered by the combination of necrosis su- and the sera amount used in this study was much pernatant and SLE sera (p<0.05) (Figure 3b). The smaller compared to that in previous reports26. blocked ICs group showed higher BAFF mRNA Given that nuclear material is released by necro- level compared to the untreated group, whereas tic cells after freeze-thawing due to a rapid loss no significant difference was found. These data of membrane integrity, and autoantibodies spe- suggested that the TNF-α and BAFF expression cific to nucleic components are present in SLE induced by the combination of necrosis superna- sera, a likely reason for the stimulatory activity tant and SLE sera in U937 cells were mediated by of the combination of necrosis supernatant and HMGB1 and RAGE. SLE sera is the formation of ICs. TNF is a strong mediator of inflammatory responses and is im- The DNA Component of ICs Contributes plicated in the tissue inflammation and damage. to Cellular Activity TNF is increased in the serum of active SLE and To test our hypothesis that the DNA compo- in renal tissue with active nephritis, and the le- nent in ICs is necessary for inducing cellular acti- vels of TNF is correlated with renal disease activation, we stimulated U937 cells with ICs, DNa- vity6,28. The present work showed that the com- se or RNase-treated ICs. We found that DNase bination of necrosis supernatant and SLE sera (at a final concentration of 20 u/ml) treatment stimulated TNF-α production from U937 cells, of ICs significantly neutralized their TNF-α and which suggests that ICs may be an important

1815 X.-J. Gao, Y.-Y. Qu, X.-W. Liu, M. Zhu, C.-Y. Ma, Y.-L. Jiao, B. Cui, Z.-J. Chen, Y.-R. Zhao

Figure 4. Effect of DNAase on the TNF-α and BAFF-inducing activity of preformed ICs. U937 cells were stimulated by ICs or ICs pretreated with the indicated concentrations of DNase for 1 h. Cells were harvested at the 8 h time point for RNA. The relative expression depicted as the extent of change with respect to GAPDH values: (a) TNF-α and (b) BAFF. factor responsible for elevated level of TNF-α in autoimmune B cells in human SLE. Notably, the the serum and renal tissue of patients with active present study indicates that ICs have the potential SLE. Also, monocytes activated by ICs may be to drive the production of BAFF from monocytes a critical source of TNF-α. TNF-α can mediate in SLE patients. Our finding provides an expla- monocyte migration and induce human endothe- nation for the ongoing BAFF production that is lial cell activation and monocyte-endothelial cell commonly found in SLE patients, and also we adhesion, which facilitates monocyte chemota- have suggested that the mechanism is to be pi- xis and deposition to tissues4. Renal pathology votal in the population expansion of autoreactive shows that lupus nephritis is characterized by B cells. In vivo high-level of BAFF may relax ICs deposition and inflammatory cell infiltra- B cell selection, which leads to the increase of tion. In this study, mononuclear cell infiltration B cells hyper-reactive to auto-antigens. These B of the kidney was observed during the initial cells produce pathological auto-antibodies com- murine lupus nephritis29. Moreover, data from monly to nucleic antigens, exacerbating SLE di- patients with lupus nephritis demonstrated that sease state20. Thus, this established effect of ICs infiltrating monocytes vigorously participate in may contribute to the induction and development renal inflammation and injury24,30. Thus, TNF-α of SLE. Several investigations suggested that over-expression of monocytes induced by ICs HMGB1 is associated with circulating DNA/ will promote the inflammatory course and aggra- nucleosomes-containing ICs from SLE patients vate tissue injury in SLE. Abnormal activation and is required for the immunostimulatory acti- of B cells and production of pathogenic autoan- vity of ICs33,34. The pro-inflammatory activity of tibodies are important contributors to the deve- HMGB1 is mostly attributed to its ligation with lopment of SLE22. BAFF is a prominent factor RAGE, which is the first described receptor for critical for B cell activation and function. BAFF HMGB1, and is expressed on monocytes, ma- deficiency results in decreased peripheral B cel- crophages and other cell types5. Many kinds of ls and a diminished ability to mount robust hu- biological effects of HMGB1 are proposed to be moral immune responses18, whereas transgenic mediated by RAGE25,26,35. Our recent report36 has murine overexpressing BAFF developed a SLE- also shown that ICs can induce endothelial cell like syndrome31. Moreover, circulating BAFF dysfunction via the HMGB1-RAGE axis in the levels are elevated and correlate positively with pathogenesis of SLE vasculitis. Here the capa- anti-dsDNA antibody titers and disease activity city of ICs to induce TNF-α and BAFF expres- in SLE patients32. These data suggest that BAFF sion was inhibited by HMGB1 antagonist A-box, is closely associated with elevated production which suggests that HMGB1 plays a critical role of anti-DNA antibodies and it is likely to play in the expression of TNF-α and BAFF. More- a critical role in the activation of antigen-driven over, the capacity of ICs to trigger TNF-α and

1816 Immune complexes induce TNF-α and BAFF production from U937 cells by HMGB1 and RAGE

BAFF production was also inhibited by anti-RA- Acknowledgements GE antibody, which further suggests that RAGE This research was funded by grants from Shandong Pro- is involved in response to ICs. Altogether, these vincial Key Technologies R&D Program (Grant No. data demonstrate that the effects of ICs on TNF-α 2009GG10002008), Shandong Provincial Natural Science and BAFF expression in U937 cells are mediated Foundation (Grant No. ZR2009CM139) and Yantai City Sci- ence & Technology Project Funds (Grant No. 2014WS002). via HMGB1 and RAGE. Our observations are We thank the participants in the study for making this work consistent with previous findings that HMGB1 possible. was critical in the pro-inflammatory activity of lupus-associated ICs26,33. These findings provide evidence that ICs can activate the immune cells Conflict of interest by HMGB1, and, thereby, may contribute to the The authors declare no conflicts of interest. loss of immunological tolerance against self-antigens. The sticky nature of HMGB1 enables it to synergize and complex with other molecules References such as DNA17,21,37. Thus, HMGB1 has been re- garded as one of the components in DNA-containing ICs3. The present finding that the DNA 1) Rahman A, Isenberg DA. Systemic lupus erythema- component was required for ICs activity, which tosus. N Engl J Med 2008; 358: 929-939. Liu YJ, Fan WJ, Bai JZ suggests that HMGB1 and DNA component are 2) . microRNA-126 expression and its mechanism of action in patients with sy- both involved in the immune-stimulatory effect stemic lupus erythematosus. Eur Rev Med Phar- of ICs in patients with SLE. Our data support macol Sci 2015; 19: 3838-3842. previous reports that ICs stimulate the activa- 3) Margiotta D, Navarini L, Vadacca M, Basta F, Lo Vul- tion and cytokine production of immune cells lo M, Pignataro F, Zardi EM, Afeltra A. Relationship via cooperative interactions among different re- between leptin and regulatory T cells in systemic ceptors, in which TLR, Fc-receptor and RAGE lupus erythematosus: preliminary results. Eur Rev are involved26,38,39. Therefore, one explanation Med Pharmacol Sci 2016; 20: 636-641. 4) Ishii KJ, Coban C, Kato H, Takahashi K, Torii Y, Ta- for the enhanced immune-stimulatory activity of keshita F, Ludwig H, Sutter G, Suzuki K, Hemmi H, the complex is that the components of ICs can Sato S, Yamamoto M, Uematsu S, Kawai T, Takeuchi in synergy contribute to cellular effects through O, Akira S. A toll-like receptor-independent antivi- the participation of different receptors. In view ral response induced by double-stranded B-form of established roles for Fc-receptor, TLR-9 and DNA. Nat Immunol 2006; 7: 40-48. RAGE in response to ICs26,38,40, the exact roles of 5) Bianchi ME, Agresti A. HMG proteins: dynamic the components in ICs as well as the downstream players in gene regulation and differentiation. mechanism remain to be further clarified. Curr Opin Genet Dev 2005; 15: 496-506. 6) Ivanov S, Dragoi AM, Wang X, Dallacosta C, Louten J, Musco G, Sitia G, Yap GS, Wan Y, Biron CA, Bian- chi ME, Wang H, Chu WM. A novel role for HMGB1 Conclusions in TLR9-mediated inflammatory responses to CpG-DNA. Blood 2007; 110: 1970-1981. Our data are in line with a model in which ne- 7) Park JS, Svetkauskaite D, He Q, Kim JY, Strassheim D, crotic cells release HMGB1. HMGB1 binds to Ishizaka A, Abraham E. Involvement of toll-like re- DNA-containing ICs in serum and then the resul- ceptors 2 and 4 in cellular activation by high mo- tant complexes stimulate TNF-α and BAFF pro- bility group box 1 protein. J Biol Chem 2004; 279: duction from monocytes by HMGB1 and RAGE. 7370-7377. Rouhiainen A, Tumova S, Valmu L, Kalkkinen N, Rau- Also, the effect observed here provides a possi- 8) vala H. Pivotal advance: analysis of proinflamma- ble mechanism for increasing BAFF production tory activity of highly purified eukaryotic recom- in SLE, whereby B cell activation, antibody pro- binant HMGB1 (amphoterin). J Leukoc Biol 2007; duction and ICs stimulated monocytes create a 81: 49-58. vicious cycle that leads to B cell hyperactivity. 9) Sha Y, Zmijewski J, Xu Z, Abraham E. HMGB1 de- Our results collectively provided evidence that velops enhanced proinflammatory activity by ICs containing DNA are implicated in monocyte binding to cytokines. J Immunol 2008; 180: 2531- disorder, which may contribute to the immune 2537. Youn JH, Oh YJ, Kim ES, Choi JE, Shin JS. dysregulation in SLE, and that interference with 10) High mobility group box 1 protein binding to lipopolysaccha- the action of HMGB1 in these ICs could be one ride facilitates transfer of lipopolysaccharide to strategy for the development of new therapies. CD14 and enhances lipopolysaccharide-media-

1817 X.-J. Gao, Y.-Y. Qu, X.-W. Liu, M. Zhu, C.-Y. Ma, Y.-L. Jiao, B. Cui, Z.-J. Chen, Y.-R. Zhao

ted TNF-alpha production in human monocytes. 23) Tian J, Avalos AM, Mao SY, Chen B, Senthil K, Wu H, J Immunol 2008; 180: 5067-5074. Parroche P, Drabic S, Golenbock D, Sirois C, Hua J, An LL, Audoly L, La Rosa G, Bierhaus A, Naworth P, 11) Urbonaviciute V, Furnrohr BG, Meister S, Munoz Marshak-Rothstein A, Crow MK, Fitzgerald KA, Latz L, Heyder P, De Marchis F, Bianchi ME, Kirschning E, Kiener PA, Coyle AJ. C, Wagner H, Manfredi AA, Kalden JR, Schett G, Toll-like receptor 9-depen- Rovere-Querini P, Herrmann M, Voll RE. Induction dent activation by DNA-containing immune com- of inflammatory and immune responses by HM- plexes is mediated by HMGB1 and RAGE. Nat GB1-nucleosome complexes: implications for Immunol 2007; 8: 487-496. the pathogenesis of SLE. J Exp Med 2008; 205: 24) Hochberg MC. Updating the American College of 3007-3018. Rheumatology revised criteria for the classifica- 12) Katsiari CG, Liossis SN, Sfikakis PP. The pathophy- tion of systemic lupus erythematosus. Arthritis siologic role of monocytes and macrophages in Rheum 1997; 40: 1725. systemic lupus erythematosus: a reappraisal. Se- 25) Li J, Wang H, Mason JM, Levine J, Yu M, Ulloa L, min Arthritis Rheum 2010; 39: 491-503. Czura CJ, Tracey KJ, Yang H. Recombinant HMGB1 13) Kyttaris VC, Juang YT, Tsokos GC. Immune cells and with cytokine-stimulating activity. J Immunol cytokines in systemic lupus erythematosus: an Methods 2004; 289: 211-223. update. Curr Opin Rheumatol 2005; 17: 518-522. 26) Livak KJ, Schmittgen TD. Analysis of relative gene expression data using Real-time quantitative PCR 14) Lourenco EV, La Cava A. Cytokines in systemic lupus and the 2(-Delta Delta C(T)) method. Methods erythematosus. Curr Mol Med 2009; 9: 242-254. 2001; 25: 402-408. 15) Lee HM, Sugino H, Nishimoto N. Cytokine networks 27) Pisetsky DS. The immune response to cell death in in systemic lupus erythematosus. J Biomed Bio- SLE. Autoimmun Rev 2004; 3: 500-504. technol 2010; 2010: 676284. Yang H, Ochani M, Li J, Qiang X, Tanovic M, Harris HE, Moore PA, Belvedere O, Orr A, Pieri K, LaFleur DW, 28) 16) Susarla SM, Ulloa L, Wang H, DiRaimo R, Czura CJ, Feng P, Soppet D, Charters M, Gentz R, Parmelee D, Li Wang H, Roth J, Warren HS, Fink MP, Fenton MJ, An- Y, Galperina O, Giri J, Roschke V, Nardelli B, Carrell dersson U, Tracey KJ J, Sosnovtseva S, Greenfield W, Ruben SM, Olsen HS, . Reversing established sepsis with antagonists of endogenous high-mobility group Fikes J, Hilbert DM. BLyS: member of the tumor ne- box 1. Proc Natl Acad Sci USA 2004; 101: 296-301. crosis factor family and B lymphocyte stimulator. Science 1999; 285: 260-263. 29) Means TK, Latz E, Hayashi F, Murali MR, Golenbock DT, Luster AD. Human lupus autoantibody-DNA 17) Cheema GS, Roschke V, Hilbert DM, Stohl W. Ele- complexes activate DCs through cooperation of vated serum B lymphocyte stimulator levels in CD32 and TLR9. J Clin Invest 2005; 115: 407-417. patients with systemic immune-based rheumatic Scaffidi P, Misteli T, Bianchi ME diseases. Arthritis Rheum 2001; 44: 1313-1319. 30) . Release of chro- matin protein HMGB1 by necrotic cells triggers 18) Schneider P, MacKay F, Steiner V, Hofmann K, Bod- mer JL, Holler N, Ambrose C, Lawton P, Bixler S, inflammation. Nature 2002; 418: 191-195. Acha-Orbea H, Valmori D, Romero P, Werner-Favre 31) Malide D, Russo P, Bendayan M. Presence of tumor C, Zubler RH, Browning JL, Tschopp J. BAFF, a novel necrosis factor alpha and interleukin-6 in renal ligand of the tumor necrosis factor family, stimu- mesangial cells of lupus nephritis patients. Hum lates B cell growth. J Exp Med 1999; 189: 1747- Pathol 1995; 26: 558-564. 1756. 32) Herrera-Esparza R, Barbosa-Cisneros O, Villalo- bos-Hurtado R, Avalos-Diaz E. 19) Petri M, Stohl W, Chatham W, McCune WJ, Chevrier Renal expression M, Ryel J, Recta V, Zhong J, Freimuth W. Associa- of IL-6 and TNFalpha genes in lupus nephritis. tion of plasma B lymphocyte stimulator levels and Lupus 1998; 7: 154-158. disease activity in systemic lupus erythematosus. 33) Rajesh M, Mukhopadhyay P, Batkai S, Haskó G, Liaudet Arthritis Rheum 2008; 58: 2453-2459. L, Huffman JW, Csiszar A, Ungvari Z, Mackie K, Chat- terjee S, Pacher P. 20) Bave U, Alm GV, Ronnblom L. The combination of CB2-receptor stimulation atte- apoptotic U937 cells and lupus IgG is a potent IFN-al- nuates TNF-alpha-induced human endothelial cell pha inducer. J Immunol 2000; 165: 3519-3526. activation, transendothelial migration of monocytes, and monocyte-endothelial adhesion. Am J Physiol Gross JA, Johnston J, Mudri S, Enselman R, Dillon 21) Heart Circ Physiol 2007; 293: H2210-2218. SR, Madden K, Xu W, Parrish-Novak J, Foster D, isetsky Lofton-Day C, Moore M, Littau A, Grossman A, Hau- 34) P DS. The complex role of DNA, histones gen H, Foley K, Blumberg H, Harrison K, Kindsvogel and HMGB1 in the pathogenesis of SLE. Autoim- W, Clegg CH. TACI and BCMA are receptors for a munity 2014; 47: 487-493. TNF homologue implicated in B-cell autoimmune 35) Perez de Lema G, Maier H, Nieto E, Vielhauer V, disease. Nature 2000; 404: 995-999. Luckow B, Mampaso F, Schlöndorff D. Chemokine 22) Lovgren T, Eloranta ML, Bave U, Alm GV, Ronnblom expression precedes inflammatory cell infiltration L. Induction of interferon-alpha production in pla- and chemokine receptor and cytokine expression smacytoid dendritic cells by immune complexes during the initiation of murine lupus nephritis. J containing nucleic acid released by necrotic or Am Soc Nephrol 2001; 12: 1369-1382. late apoptotic cells and lupus IgG. Arthritis Rheum 36) Sun W, Jiao Y, Cui B, Gao X, Xia Y, Zhao Y. Immune 2004; 50: 1861-1872. complexes activate human endothelium involving

1818 Immune complexes induce TNF-α and BAFF production from U937 cells by HMGB1 and RAGE

the cell-signaling HMGB1-RAGE axis in the pa- 39) Ikezumi Y, Suzuki T, Hayafuji S, Okubo S, Nikolic-Pa- thogenesis of lupus vasculitis. Lab Invest 2013; terson DJ, Kawachi H, Shimizu F, Uchiyama M. The 93: 626-638. sialoadhesin (CD169) expressing a macropha- 37) Yang N, Isbel NM, Nikolic-Paterson DJ, Li Y, Ye R, ge subset in human proliferative glomerulone- Atkins RC, Lan HY. Local macrophage proliferation phritis. Nephrol Dial Transplant 2005; 20: 2704- in human glomerulonephritis. Kidney Int 1998; 54: 2713. 143-151. 40) Nirala BK, Gohil NK. Effect of garlic component 38) Cancro MP, D’Cruz DP, Khamashta MA. The role of s-allyl cysteine sulfoxide on glycated human se- B lymphocyte stimulator (BLyS) in systemic lupus rum albumin induced activation of endothelial cel- erythematosus. J Clin Invest 2009; 119: 1066- ls: an in vitro study. Eur Rev Med Pharmacol Sci 1073. 2015; 19: 2125-2131.

1819